Ethylene is Involved in Symptom Development and Ribosomal Stress of Tomato Plants upon Citrus Exocortis Viroid Infection

Citrus exocortis viroid (CEVd) is known to cause different symptoms in citrus trees, and its mechanism of infection has been studied in tomato as an experimental host, producing ribosomal stress on these plants. Some of the symptoms caused by CEVd in tomato plants resemble those produced by the phytohormone ethylene. The present study is focused on elucidating the relationship between CEVd infection and ethylene on disease development. To this purpose, the ethylene insensitive Never ripe (Nr) tomato mutants were infected with CEVd, and several aspects such as susceptibility to infection, defensive response, ethylene biosynthesis and ribosomal stress were studied. Phenotypic characterization revealed higher susceptibility to CEVd in these mutants, which correlated with higher expression levels of both defense and ethylene biosynthesis genes, as well as the ribosomal stress marker SlNAC082. In addition, Northern blotting revealed compromised ribosome biogenesis in all CEVd infected plants, particularly in Nr mutants. Our results indicate a higher ethylene biosynthesis in Nr mutants and suggest an important role of this phytohormone in disease development and ribosomal stress caused by viroid infection.


Introduction
Viroids are the smallest known plant pathogens and consist of circular, highly structured, non-coding RNA molecules with autonomous replication that parasite the transcriptional machinery of their hosts [1,2]. Viroids can be classified into two families, the Avsunviroidae, whose members replicate in the chloroplasts and have a hammerhead-like structure, and the Pospiviroidae, characterized by a rod-like structure and replicating in the nucleus [3]. Citrus exocortis viroid (CEVd), which belongs to the Pospiviroidae family, consists of around 370 nucleotides and has a broad range of hosts including tomato plants (Solanum lycopersicum L.). CEVd symptoms in tomato include stunting, epinasty, midvein necrosis, chlorosis and leaf rugosity [4][5][6]. Some of these symptoms have been associated with those produced by ethylene, since plants exogenously treated with the ethylene-releasing compound ethephon, as well as mutants that over produce ethylene, display similar symptoms to those caused by CEVd [7][8][9]. However, the relationship between ethylene production and signaling, and the disease caused by CEVd, has not been deeply studied.
Ethylene (ET) is a small phytohormone involved in plant growth, development and stress response. Despite its low basal levels, a quick induction of its biosynthesis can be observed under biotic or abiotic stress and senescence [10]. Due to its gaseous nature, ethylene can easily move in the plant without To better quantify the observed Nr hyper-susceptibility, the percentage of plants displaying mild epinasty was tracked to obtain a graphic representation of disease development ( Figure 2). The accelerated appearance of symptoms was observed in Nr mutants with 87.5% plants showing symptoms at 12 days post inoculation (dpi), when only 50% parental plants exhibited them. At 14 dpi, all the Nr mutant plants displayed symptoms, while 12.5% Rutgers plants remained symptomless. These results suggest an accelerated symptom appearance in Nr mutant tomato plants, confirming the higher susceptibility observed.  To better quantify the observed Nr hyper-susceptibility, the percentage of plants displaying mild epinasty was tracked to obtain a graphic representation of disease development ( Figure 2). The accelerated appearance of symptoms was observed in Nr mutants with 87.5% plants showing symptoms at 12 days post inoculation (dpi), when only 50% parental plants exhibited them. At 14 dpi, all the Nr mutant plants displayed symptoms, while 12.5% Rutgers plants remained symptomless. These results suggest an accelerated symptom appearance in Nr mutant tomato plants, confirming the higher susceptibility observed.
Plants 2020, 9,582 3 of 16 with Nr mutants displaying a much more severe epinasty, stunting and leaf rugosity, and therefore appearing to be more susceptible to the viroid infection. To better quantify the observed Nr hyper-susceptibility, the percentage of plants displaying mild epinasty was tracked to obtain a graphic representation of disease development ( Figure 2). The accelerated appearance of symptoms was observed in Nr mutants with 87.5% plants showing symptoms at 12 days post inoculation (dpi), when only 50% parental plants exhibited them. At 14 dpi, all the Nr mutant plants displayed symptoms, while 12.5% Rutgers plants remained symptomless. These results suggest an accelerated symptom appearance in Nr mutant tomato plants, confirming the higher susceptibility observed.   To verify the statistical significance of the differences in the disease development shown in Figure 2, a scale of the disease severity was developed, scoring symptoms from mild (mild epinasty) to very severe (midvein necrosis and chlorosis), at different time points (see Materials and Methods). At the beginning of the viroid infection (12 dpi), 12.5% of the Nr mutants already showed very severe symptoms, whilst Rutgers plants remained symptomless or displayed mild symptoms ( Figure 3). These differences continued throughout the experiment, with most Nr plants (87.5%) showing very severe symptoms by 19 dpi. Severe symptomatology was not observed in Rutgers plants until 21 dpi, and by the end of the experiment, all of the Nr plants, but only 12.5% of Rutgers plants showed severe symptoms. In conclusion, differences in symptom severity were statistically significant at any time between Nr and Rutgers tomato plants, correlating with symptom appearance, and thus confirming the hyper-susceptibility of Nr tomato mutants to CEVd infection. To verify the statistical significance of the differences in the disease development shown in Figure 2, a scale of the disease severity was developed, scoring symptoms from mild (mild epinasty) to very severe (midvein necrosis and chlorosis), at different time points (see Materials and Methods). At the beginning of the viroid infection (12 dpi), 12.5% of the Nr mutants already showed very severe symptoms, whilst Rutgers plants remained symptomless or displayed mild symptoms ( Figure 3). These differences continued throughout the experiment, with most Nr plants (87.5%) showing very severe symptoms by 19 dpi. Severe symptomatology was not observed in Rutgers plants until 21 dpi, and by the end of the experiment, all of the Nr plants, but only 12.5% of Rutgers plants showed severe symptoms. In conclusion, differences in symptom severity were statistically significant at any time between Nr and Rutgers tomato plants, correlating with symptom appearance, and thus confirming the hyper-susceptibility of Nr tomato mutants to CEVd infection. . Disease severity of CEVd infected Rutgers and Never ripe tomato plants. Symptomatology was scored at 12, 14, 19, 21 and 26 days post inoculation (dpi), using the following scale: no symptoms (white), mild epinasty (light grey), severe epinasty and stunting (grey), leaf rugosity (dark grey), midvein necrosis and chlorosis (black). Data correspond to one representative experiment. Data were analyzed using a Mann-Whitney test and different letters indicate significant differences (p < 0.05). Number of (') indicates different days.
To study the possible relationship between the observed symptomatology and pathogen accumulation, CEVd levels were analyzed in Rutgers and Nr infected tomato plants at 4 wpi by qRT-PCR (see Materials and Methods). Surprisingly, a higher accumulation of CEVd was observed in infected parental Rutgers plants at 4 wpi ( Figure 4), when compared to the hyper-susceptible Nr mutants. These data indicate that symptom development appears not to be correlated to the accumulation of CEVd. . Disease severity of CEVd infected Rutgers and Never ripe tomato plants. Symptomatology was scored at 12, 14, 19, 21 and 26 days post inoculation (dpi), using the following scale: no symptoms (white), mild epinasty (light grey), severe epinasty and stunting (grey), leaf rugosity (dark grey), midvein necrosis and chlorosis (black). Data correspond to one representative experiment. Data were analyzed using a Mann-Whitney test and different letters indicate significant differences (p < 0.05). Number of (') indicates different days.
To study the possible relationship between the observed symptomatology and pathogen accumulation, CEVd levels were analyzed in Rutgers and Nr infected tomato plants at 4 wpi by qRT-PCR (see Materials and Methods). Surprisingly, a higher accumulation of CEVd was observed in infected parental Rutgers plants at 4 wpi ( Figure 4), when compared to the hyper-susceptible Nr mutants. These data indicate that symptom development appears not to be correlated to the accumulation of CEVd.

PR1 and ACS2 Are Highly Induced in Infected Nr Tomato Plants
To find out if Nr mutants, which have blocked ethylene perception, were also impaired in the activation of the defensive response against CEVd, the accumulation and expression of pathogenesis related protein 1 (PR1; accession X71592) [39], which has been described as a classical marker of plant defense that is rapidly induced in CEVd-infected tomato plants [40,41]. Besides, the induction of genes involved in ethylene biosynthesis was also analyzed.
For the analysis of defense proteins, both PR1 accumulation and PR1 gene expression were measured in control and CEVd-infected tomato plants of both genotypes (see Materials and Methods), and data were statistically analyzed ( Figure 5). As expected, accumulation of PR1 protein increased over time in all plants as the disease progressed, its high levels being even detectable by Coomassie Blue stain (Figure 5a). Interestingly, a higher accumulation of PR1 was observed for all time points in infected Nr mutants when compared to their parental Rutgers plants. Expression levels of PR1 gene correlated with PR1 protein levels and the hyper-susceptibility observed in Nr mutants (Figure 5b). In fact, the induction of PR1 was statistically higher in the infected Nr mutants at 3 and 4 wpi than in the wild type Rutgers.
To study the possible relationship between the observed symptomatology and pathogen accumulation, CEVd levels were analyzed in Rutgers and Nr infected tomato plants at 4 wpi by qRT-PCR (see Materials and Methods). Surprisingly, a higher accumulation of CEVd was observed in infected parental Rutgers plants at 4 wpi (Figure 4), when compared to the hyper-susceptible Nr mutants. These data indicate that symptom development appears not to be correlated to the accumulation of CEVd.  . Data correspond to the mean of two or more independent plants ± SD of at least 3 technical replicates. Data displayed correspond to one representative experiment. Data were analyzed using a Mann-Whitney test and different letters indicate significant differences (p < 0.05).

PR1 and ACS2 Are Highly Induced in Infected Nr Tomato Plants
To find out if Nr mutants, which have blocked ethylene perception, were also impaired in the activation of the defensive response against CEVd, the accumulation and expression of pathogenesis related protein 1 (PR1; accession X71592) [39], which has been described as a classical marker of plant defense that is rapidly induced in CEVd-infected tomato plants [40,41]. Besides, the induction of genes involved in ethylene biosynthesis was also analyzed.
For the analysis of defense proteins, both PR1 accumulation and PR1 gene expression were measured in control and CEVd-infected tomato plants of both genotypes (see Materials and Methods), and data were statistically analyzed ( Figure 5). As expected, accumulation of PR1 protein increased over time in all plants as the disease progressed, its high levels being even detectable by Coomassie Blue stain ( Figure 5A). Interestingly, a higher accumulation of PR1 was observed for all time points in infected Nr mutants when compared to their parental Rutgers plants. Expression levels of PR1 gene correlated with PR1 protein levels and the hyper-susceptibility observed in Nr mutants ( Figure 5B). In fact, the induction of PR1 was statistically higher in the infected Nr mutants at 3 and 4 wpi than in the wild type Rutgers. The contribution of ethylene in Rutgers tomato plants after CEVd infection has already been described [5,15]. However, less is known about the role of ethylene in CEVd-infected Nr mutants. To analyze the ethylene implication in the development of CEVd symptoms in tomato plants, the expression levels of 1-aminocyclopropane-1-carboxylate synthase 2 (ACS2; accession X59145.1) ( Figure  6A) and 1-aminocyclopropane-1-carboxylate oxidase (ACO1; accession X58273.1) ( Figure 6B) genes were The contribution of ethylene in Rutgers tomato plants after CEVd infection has already been described [5,15]. However, less is known about the role of ethylene in CEVd-infected Nr mutants. To analyze the ethylene implication in the development of CEVd symptoms in tomato plants, the expression levels of 1-aminocyclopropane-1-carboxylate synthase 2 (ACS2; accession X59145.1) ( Figure 6a) and 1-aminocyclopropane-1-carboxylate oxidase (ACO1; accession X58273.1) (Figure 6b) genes were analyzed, as the induction of these isoforms has been described upon pathogen attack, More interestingly, these levels were always statistically higher in infected Nr mutants when compared to the corresponding infected Rutgers plants. Similarly, ACO1 expression levels ( Figure 6b) were also higher in CEVd infected plants when compared to the non-infected control plants, but showed no significant differences between infected Nr mutants and Rutgers plants.
Plants 2020, 9,582 6 of 16 analyzed, as the induction of these isoforms has been described upon pathogen attack, and at the onset of the climacteric stage in tomato fruits, where a high accumulation of ethylene is produced [7, 42,43]. The expression levels of ACS2 ( Figure 6A) were higher in CEVd-infected plants when compared with the corresponding non-infected plants, at any time point and for both genotypes. More interestingly, these levels were always statistically higher in infected Nr mutants when compared to the corresponding infected Rutgers plants. Similarly, ACO1 expression levels ( Figure  6B) were also higher in CEVd infected plants when compared to the non-infected control plants, but showed no significant differences between infected Nr mutants and Rutgers plants. Our results indicate a distinctive increase in both PR1 accumulation and ethylene biosynthesis gene expression during CEVd infection, and this induction is enhanced in Nr mutants.

Ethylene Production Is Increased in Nr Tomato Mutants upon CEVd Infection
Ethylene levels were measured to correlate the observed increase in the expression of ethylene biosynthesis genes in Nr infected plants with the emitted ethylene ( Figure 7). In Rutgers tomato plants, no significant differences in ethylene levels were observed between control and CEVd infected plants at any time. In contrast, Nr infected mutants showed higher ethylene levels at any time point, which correlates with the enhanced expression levels of ACS2 ( Figure 6A).
Our results indicate that ethylene levels do not correlate with ACS2 and ACO1 gene expression in Rutgers plants, but closely correlate in Nr mutants. Our results indicate a distinctive increase in both PR1 accumulation and ethylene biosynthesis gene expression during CEVd infection, and this induction is enhanced in Nr mutants.

Ethylene Production Is Increased in Nr Tomato Mutants upon CEVd Infection
Ethylene levels were measured to correlate the observed increase in the expression of ethylene biosynthesis genes in Nr infected plants with the emitted ethylene (Figure 7). In Rutgers tomato plants, no significant differences in ethylene levels were observed between control and CEVd infected plants at any time. In contrast, Nr infected mutants showed higher ethylene levels at any time point, which correlates with the enhanced expression levels of ACS2 (Figure 6a).
Our results indicate that ethylene levels do not correlate with ACS2 and ACO1 gene expression in Rutgers plants, but closely correlate in Nr mutants.

Ribosomal Stress Is Enhanced in Never Ripe Tomato Mutants upon CEVd Infection
We previously reported that CEVd produces ribosomal stress in tomato plants [30]. To explore the possible role of ethylene in the ribosomal stress caused by viroids, SlNAC082 (accession Solyc11g005920.

Ribosomal Stress Is Enhanced in Never Ripe Tomato Mutants upon CEVd Infection
We previously reported that CEVd produces ribosomal stress in tomato plants [30]. To explore the possible role of ethylene in the ribosomal stress caused by viroids, SlNAC082 (accession Solyc11g005920.    . Ethylene emission levels in both mock and CEVd infected Rutgers and Nr tomato leaves at two, three, and four weeks post inoculation (wpi). Data correspond to the mean ± SD of at least two biological replicates. Results are shown from one representative experiment. Data were analyzed using a Mann-Whitney test and different letters indicate significant differences (p < 0.05). Number of (') indicates different weeks.

Ribosomal Stress Is Enhanced in Never Ripe Tomato Mutants upon CEVd Infection
We previously reported that CEVd produces ribosomal stress in tomato plants [30]. To explore the possible role of ethylene in the ribosomal stress caused by viroids, SlNAC082 (accession Solyc11g005920.   To investigate whether rRNA processing is further affected in Nr mutants, a Northern blot analysis was performed with a probe targeting the P'-A3 pre-rRNA (see Materials and Methods) ( Figure 9a). As previously described [30], an overaccumulation of 35S pre-rRNA and P'-A3 was observed in CEVd-infected plants. Results were quantified by optical density analysis and a statistically higher accumulation of P'-A3 was observed in Nr mutants compared to Rutgers plants (Figure 9b). These To investigate whether rRNA processing is further affected in Nr mutants, a Northern blot analysis was performed with a probe targeting the P'-A3 pre-rRNA (see Materials and Methods) ( Figure 9A). As previously described [30], an overaccumulation of 35S pre-rRNA and P'-A3 was observed in CEVd-infected plants. Results were quantified by optical density analysis and a statistically higher accumulation of P'-A3 was observed in Nr mutants compared to Rutgers plants ( Figure 9B). These results suggest a defect in ribosome processing during CEVd infection, with a greater impact in Nr mutant plants. In conclusion, we have observed that Nr mutants display hyper-susceptibility to CEVd infection, displaying a higher induction of PR1, an increase in the activation of the ethylene biosynthetic pathway, and enhanced ribosomal stress, thus indicating that ethylene plays a key role in mediating disease development upon CEVd infection in tomato plants.

Discussion
The involvement of ethylene in viroid disease has already been described for CEVd [15] and other viroids [22,44], although its precise role has not been deciphered yet. Previous studies have shown the capacity of ethylene to provoke similar symptoms to those caused by viroid infection [9,45]. On the other hand, recent studies have revealed some correlation between CEVd presence, ribosomal stress and symptom development [30]. However, the relationship between ribosomal stress and ethylene production had not yet been clarified. Hence, the goal of this study was to explore the role of ethylene in disease development and ribosomal stress upon CEVd infection.
The ethylene insensitive Never ripe tomato mutants, which constitutively block ethylene response, were used to investigate the role of this phytohormone in symptom development. The importance of ethylene in the plant defensive response to different pathogens has been studied using these mutants, which showed less susceptibility to diseases caused by bacteria [24,46] but higher susceptibility to symptoms caused by fungi [7,23]. Our results suggest that Never ripe mutants In conclusion, we have observed that Nr mutants display hyper-susceptibility to CEVd infection, displaying a higher induction of PR1, an increase in the activation of the ethylene biosynthetic pathway, and enhanced ribosomal stress, thus indicating that ethylene plays a key role in mediating disease development upon CEVd infection in tomato plants.

Discussion
The involvement of ethylene in viroid disease has already been described for CEVd [15] and other viroids [22,44], although its precise role has not been deciphered yet. Previous studies have shown the capacity of ethylene to provoke similar symptoms to those caused by viroid infection [9,45]. On the other hand, recent studies have revealed some correlation between CEVd presence, ribosomal stress and symptom development [30]. However, the relationship between ribosomal stress and ethylene production had not yet been clarified. Hence, the goal of this study was to explore the role of ethylene in disease development and ribosomal stress upon CEVd infection.
The ethylene insensitive Never ripe tomato mutants, which constitutively block ethylene response, were used to investigate the role of this phytohormone in symptom development. The importance of ethylene in the plant defensive response to different pathogens has been studied using these mutants, which showed less susceptibility to diseases caused by bacteria [24,46] but higher susceptibility to symptoms caused by fungi [7,23]. Our results suggest that Never ripe mutants are more susceptible to CEVd infection than their corresponding wild type plants (Figure 1), supported by the accelerated appearance of symptoms (Figure 2), the greater severity of developed symptoms (Figure 3), and the higher expression and accumulation of PR1 ( Figure 5). This susceptibility seems to be specific to CEVd, since Never ripe plants infected with TCDVd have been reported to display slightly reduced symptoms [22], thus indicating differences in pathogenicity between members of the Pospiviroidae family. Because ethylene perception is impaired in these mutants, our results suggest a role of ethylene signaling in the defensive response against CEVd.
Contrary to the hyper-susceptibility observed in Never ripe mutants, CEVd accumulation was higher in infected parental plants (Figure 4), suggesting that the symptomatology may not be associated with pathogen levels, as described by other authors [5,47]. This could be explained by a faster weakening of infected Never ripe plants, which displayed enhanced ethylene-related symptoms, therefore affecting the replication capabilities of CEVd.
The analysis of ethylene production showed higher ethylene levels in CEVd infected Nr mutants when compared to wild type plants (Figure 7), which correlates with higher ACS2 expression observed in these mutants ( Figure 6). This indicates an over-activation of the ET biosynthetic pathway upon CEVd infection in mutant plants. Moreover, negative feedback regulation has been described for the biosynthesis of ethylene through the regulation of ACS expression [11,24]. Higher ACS2 and ethylene levels in Never ripe mutants could be explained by an impairment in this feedback regulation, due to the lack of signaling. ACS2 is a key enzyme in the ethylene biosynthetic pathway that is specifically induced upon necrotrophic pathogen attack [13,[48][49][50][51][52][53]. Regarding viroids, an induction in ethylene-related genes during PSTVd infection has also been described [44], which correlates with our observed results upon CEVd infection. In NahG tomato plants, which are unable to accumulate salicylic acid (SA), a dramatic increment in ethylene synthesis also occurred after CEVd inoculation [5], which could be explained by the described antagonism between the ET/JA and SA signaling pathways [7, 46,54]. In accordance with our results, these NahG tomato plants displayed a positive correlation between ethylene levels and symptom development upon CEVd infection [5], which has also been reported in other plant-pathogen interactions [8]. Besides, the viroidal symptoms can be mimicked by exogenous treatments with the ethylene-releasing agent ethephon, thus suggesting a role of ethylene in disease development [9,45]. It has also been described that, although high ethylene levels contribute to symptom development, low levels of this hormone can prevent viroid infection [55]. Our results indicate that symptom development upon CEVd infection might be caused by ethylene accumulation, even when ethylene signaling is not occurring. Here, we propose that ethylene plays a dual role in defensive response, in which its signaling is necessary for the plant defense against the viroid, and its accumulation is associated with symptom development.
This dual role has also been proposed for salicylic acid during pathogen attack. According to that model, NPR1 would act as a key regulator in SA signaling and would be degraded by NPR4 when no SA is present and by NPR3 under high SA levels [56,57]. Similarly, ethylene concentration might elicit a different response severity to the pathogen. In fact, a structural analysis of ethylene response 1 (ETR1) in Arabidopsis revealed the possibility of several ethylene binding sites for each ethylene receptor dimer [58]. We suggest a dual role of ethylene, by which low ethylene levels could contribute to slowing disease progression, but high levels might exert a toxic effect to the plant even when ethylene signaling is impaired, thus indicating an alternate route for symptom development. This is supported by reports in which Never ripe mutants from different cultivars retain slight ethylene sensitivity only under high ethylene concentrations [23,59,60]. Ethylene receptors work in homodimers [61,62] and have also been shown to form heterodimers [63][64][65]. ETR1 structural analysis showed that each monomer in the dimer may bind ethylene separately. It also reported an extra binding site between both monomers in the event of high ethylene concentrations [58]. Even though the main binding site of the NR receptor is impaired in Nr mutants, ethylene binding to the other monomer in heterodimers could explain the conserved sensitivity. At the same time, ethylene binding between monomers occurring under high ethylene concentrations may activate an alternate signaling route and indicate the possibility of ethylene binding to putative binding sites under high concentrations.
Ribosomal stress upon CEVd infection had already been described in previous studies [30]. However, no information on the relationship between this stress and ethylene had yet been reported. Here, we have detected that ET-insensitive Never ripe tomato mutants displayed higher levels of the ribosomal stress marker gene SlNAC082 [31] once the disease was sufficiently advanced (Figure 8), although no differences were observed between Nr and wild type plants for earlier infection times. These results appear to indicate that differences in SlNAC082 expression between Nr and wild type infected plants mainly occur at the late stages of infection, when a strong symptomatology is established. ANAC082 is also involved in senescence in Arabidopsis [66], explaining the increase over time of SlNAC082 expression levels in non-infected plants. Besides, SlNAC082 induction correlates as well with higher ACS2 expression and higher ethylene levels, suggesting a role of ethylene in ribosomal stress.
We have also studied whether ribosomal stress was due to defects in rRNA processing. In ribosome biogenesis, three of the rRNAs (18S, 5.8S and 25S) are processed from a single primary transcript consisting of a 5 -external transcribed spacer (5 -ETS), the 18S sequence, an internal transcribed spacer (ITS1), the 5.8S sequence, ITS2, the 25S sequence and a 3 -ETS [31,33]. Previous research showed a defect in 18S processing upon CEVd infection [30], which led us to analyze the accumulation of the immature pre-rRNA 35S and the intermediate P'-A3 ( Figure 9). As expected, a higher accumulation of both 35S and P'-A3 was observed in CEVd-infected plants when compared to non-infected plants. Moreover, Never ripe mutants also displayed a higher accumulation of both intermediates, correlating with the higher ribosomal stress observed at the same time point (Figure 8). Our results confirm that CEVd symptomatology correlates with a defect in the processing of the 18S rRNA. This effect is higher in ethylene-insensitive mutants than in wild type plants, despite their lower CEVd accumulation. Our results suggest that ethylene signaling might be necessary to alleviate the ribosomal stress caused by CEVd. On the other hand, the correlation between higher SlNAC082 expression and higher ethylene levels could also point to ethylene accumulation contributing to ribosomal stress, due to the dual role of ethylene previously proposed.
In conclusion, our results reveal the relevance of ET against the infection caused by CEVd, since ethylene insensitive Never ripe tomato mutants, which overproduce ethylene, display more severe symptoms compared to their wild type. This is not common to all viroids, indicating specific pathogenicity in CEVd. Besides, Nr plants also exhibit enhanced ribosomal stress, due to alterations in the 18S rRNA processing caused by CEVd despite the lower accumulation of viroid transcript, suggesting additional causes for the defects in pre-rRNA processing observed, and indicating an implication of ethylene in the defense against this stress caused by CEVd. We propose a dual role of ethylene in defensive response, by which low ethylene levels could delay plant disease development and mitigate the ribosomal stress caused by CEVd, while high ethylene levels may contribute to symptom severity and ribosomal stress.

Plant Material and Viroid Inoculation
Seeds of tomato (Solanum lycopersicum L.) cultivar Rutgers and the mutant Never ripe (Nr) were obtained from the Tomato Genetics Resource Center, UC Davis (https://tgrc.ucdavis.edu; accessions LA3001 and LA1090, respectively). Seeds were sterilized with a 1:1 mixture of commercial sodium hypochlorite and distilled H 2 O, and plants were grown in pots with a mixture of vermiculite and peat (1:1), which were irrigated with Hoagland solution.
A total of 35 Rutgers and Never ripe plants were used for each experiment. Plants were cultivated in a growth chamber with a 16 h light and 8 h darkness photoperiod and a temperature and relative humidity range of 28 • C/24 • C and 60%/85% (day/night), respectively. Viroidal inoculum was prepared from leaves of CEVd infected Rutgers tomato plants as previously described [67]. Ten plants were mock-inoculated with water and the rest were infected with CEVd (accession S67446) by inoculating with carborundum the first cotyledon and the first leaf of 2-week-old plants [5]. The apex and the two youngest leaves were sampled for all measurements throughout the experiment. Plants were inspected and symptom severity was scored at 1.5, 2, 2.5, 3 and 3.5 weeks post-inoculation (wpi) using the following scale: no symptoms (0), mild epinasty (1), severe epinasty and stunting (2), leaf rugosity (3), midvein necrosis and chlorosis (4).

Ethylene Measurements
Rutgers and Nr leaflets (0.5 g) showing symptoms of CEVd infection were harvested as previously described [5]. Samples were placed in the growth chamber for 4 h inside 10-mL glass vials sealed with a rubber septum, and 400 µL of the gas phase was analyzed. A 4890A Hewlett Packard gas chromatograph fitted with a flame ionization detector (FID) with a Teknokroma capillary column (2 m × 1/6" OD × 1 mm ID, Alumina F1 80/100) was used for ethylene measurements. Helium was used as the carrier gas with a pressure of 140 kPa. Injector and detector temperature were set at 200 • C and oven temperature was set at 80 • C. The ethylene peak retention time under these conditions was 2.5 min. Three replicates were measured for each time point and recorded data were analyzed with the Masslynx Waters software, using an ethylene standard curve.

RNA Preparation
Total leaf RNA was extracted using TRIzol reagent (Invitrogen, Carlsbad, CA, USA) and following the manufacturer's protocol. RNA used for RT-PCR and qRT-PCR analysis was precipitated using one volume of 6 M LiCl and incubated for 3 h at 4 • C. RNA was recovered by centrifugation for 10 min at 12000 rpm and cleaned with 3 M LiCl. RNA was dissolved in DEPC water and measured using a ND-1000 Nanodrop. Concentration was adjusted to 1 µg/µL and DNA contamination was eliminated using the TURBO DNAse kit (Ambion, Austin, TX, USA), according to the manufacturer's protocol.

RT-PCR and qRT-PCR
cDNA was synthesized from 1 µg of the extracted RNA using the PrimeScript RT kit (PerfectReal Time, Takara Bio Inc., Otsu, Shiga, Japan) following the manufacturer's protocol, using the (dT) 18 and random primers. A volume of 25 µL was used for RT-PCR using 1 µL cDNA, 1 µL of each primer, 3 µL dNTPs 2.5 mM, 0.5 µL DNa polymerase and 2.5 µL of its reaction buffer 10X (Netzyme, NEED, Valencia, Spain). Reactions were carried out using a GeneAmp PCR System 2400 (Perkin Elmer, Norwalk, CT, USA) thermocycler, using the following conditions: 1 min at 94 • C followed by 30 cycles of 1 min at 94 • C, 1 min at 56.35 • C, 1 min at 72 • C and a final extension step for 5 min at 72 • C.
Quantitative qRT-PCR was carried out as previously described [68] in a 10 µL volume, using MicroAmpFast 96-Well ReactionPlate (Applied Biosystems, Foster City, CA, USA) plates and PyroTaq EvaGreen qPCR Master Mix (CMB, Madrid, Spain) in a 7500 Fast (Life Technologies, Singapore, Singapore). Actin was used as the endogenous gene of reference. Primers used are shown in Table 1. Table 1. Primer sequences used for real-time quantitative PCR.

Northern Blot Hybridization
To detect P'-A3 in total RNAs preparations, 15 µg of RNA extracted from 4 wpi plants were denatured at 65 • C for 15 min, using 4 volumes of sample buffer (50% formamide, 6% formaldehyde in 200 mM MOPS, 50 mM sodium acetate and 10 mM EDTA, pH 7.0) and separated in a 1.5% agarose gel. Equal sample loading was checked by ethydium bromide staining and UV visualization. Transferring of RNA to Nytran membranes and hybridization was performed as previously described [69]. The p2 probe (5 -GAGCGCGGCAGTCATTCGCAAGGAGCATTC-3 ) was labelled by using polynucleotide kinase and [γ-32 P]-ATP. Membranes were exposed to X-ray film and optical density corresponding to three independent repetitions was analyzed by using the ImageJ software.

Protein Extraction and Electrophoresis Analysis
Proteins were extracted from tomato leaf tissues infected with CEVd and mock-inoculated plants.
To this purpose, 0.3 g leaf tissue was homogenized in 1 mL of 50 mM Tris-Hcl (pH 7.5) containing 15 mM 2-mercaptoethanol, transferred to 1.5 mL tubes and centrifuged at 12000 rpm and 4 • C for 10 min. After this, 500 µL supernatant were transferred to a new 1.5 mL tube, then 250 µL of 3X SDS/PAGE loading buffer were added and the mixture was boiled at 95 • C for 7 min. Thirty-five µL of each sample, along with a molecular weight marker (PageRuler, Fermentas, Burlington, ON, Canada), were run in a 14% polyacrylamide gel, as previously described [70]. Gels were stained with Coomassie Brilliant Blue R-250 (Sigma-Aldrich, Chesnes, France), prepared at 0.05% in 10% acetic acid and 20% isopropanol.

Statistical Analysis
IBM SPSS Statistics 25 software was used for all statistical analysis. A p value < 0.05 was considered as statistically significant. The Shapiro-Wilk test was used for sample normality. The Mann-Whitney test was used to compare two independent non-parametric samples. A multiple group non-parametric comparison was performed by using a Kruskal-Wallis test.